On Tue, Jul 14, 2009 at 6:27 PM, Gürkan Sengün <gur...@phys.ethz.ch> wrote:
> Hello Bhavani > > Can you tell me what this RFE Compile for 128 bit patch really is good for? > Please have a look at the answer of xmp upstream. > > -------- Original Message -------- > Subject: Re: xmp 2.7.0 > Date: Tue, 14 Jul 2009 09:34:14 -0300 > From: Claudio Matsuoka <cmatsu...@gmail.com> > To: Gürkan Sengün <gur...@phys.ethz.ch> > References: <4a5c3e79.3070...@phys.ethz.ch> > > On Tue, Jul 14, 2009 at 5:14 AM, Gürkan Sengün<gur...@phys.ethz.ch> wrote: > > I am the maintainer of xmp on Debian, I also use it on Mac OS X (iMac G4 >> PowerPC). Did you see these bug reports at Debian, and could you fix them >> in >> a later tarball release? >> > > Hi Gürkan, > > I wasn't aware of these reports, thanks for the note. A few notes on > the reports: > > #483389 xmp should not expect valid input / signedness problems: will > apply the proposed patches for 2.7.1. It's generally very hard to > cover all possibilities arising from corrupted files, so further tests > will be added when needed. > > #536740 xmp: RFE Compile for 128 bit: this is really not a matter of > bits -- strictly in terms of mixer resolution it's always performed in > 32 bits, and changing the maximum number of active voices from 64 to > 128 would increase final mixer resolution from 22 to 23 bits (16 + > log2(voices)). I've never seen a module needing more than 64 virtual > channels, and 128 seems to be quite exaggerated. I'll investigate the > need to increase the number of voices for the 2.8 series. With the > current mixer values, if a module (through NNA or similar) needs more > than 64 channels, it will allocate it according to IT rules. If the > limit is increased, there can be some output distortion. > > #442147 xmp: manpage has two explanations for -l. Oops. Will fix to > make everything consistent in 2.7.1. > > #8510 XMP should reload samples after exit. This is a bit > controversial. Any other program does this? Shouldn't the MIDI player > program reload samples before playing to ensure the card is in correct > state before playing? I'm unconvinced in this case, unless this > behavior is common policy I tend to suggest closing it as wontfix or > similar. > > #191109 xmp: show time informations on verbose command line: It can be > done, I'll check what cplay needs. > > > Will there ever be an xxmp? >> > > I don't see it coming back because the player plugins fill this gap > for users needing a GUI. > > > Do you have a powerpc machine to test hear songs there? I got a few songs >> which sound wrong on my mac (even with 2.7.0), and for some songs (from >> keygenmusic webpage) xmp just keeps allocating memory until the machine >> goes >> out of memory, >> would it help to have copies of those songs, or a shell account on my >> machine? >> > > Ugh. I'll see if I find a big-endian machine around for tests, but in > the meantime if you send me some faulty modules I can check the code. > Sometimes it's easy to find even without access to the host where the > problems happen, so Iet's try it first and if I didn't find anything > I'll ask for the account. > > Best regards, > Claudio > Hi maintainer We in ubuntu have updated our package to the latest upstream version and looking at upstream comment .. Bug is redundant I think Regards -- Bhavani Shankar.R https://launchpad.net/~bhavi, a proud ubuntu community member. What matters in life is application of mind!, It makes great sense to have some common sense..!
